README.md

PanJu offset detect by image

Use fintune from google/vit-base-patch16-224(https://huggingface.co/google/vit-base-patch16-224)

Dataset

DatasetDict({
    train: Dataset({
        features: ['image', 'label'],
        num_rows: 329
    })
    validation: Dataset({
        features: ['image', 'label'],
        num_rows: 56
    })
})

36 Break and 293 Normal in train 5 Break and 51 Normal in validation

Intended uses

How to use

Here is how to use this model to classify an image of the COCO 2017 dataset into one of the 1,000 ImageNet classes:

# Load image
import torch
from transformers import ViTFeatureExtractor, ViTForImageClassification,AutoModel
from PIL import Image
import requests
url='https://datasets-server.huggingface.co/assets/ShihTing/IsCausewayOffset/--/ShihTing--IsCausewayOffset/validation/0/image/image.jpg'
image = Image.open(requests.get(url, stream=True).raw)
# Load model
from transformers import AutoFeatureExtractor, AutoModelForImageClassification
device = torch.device('cpu')
extractor = AutoFeatureExtractor.from_pretrained('ShihTing/PanJuOffset_TwoClass')
model = AutoModelForImageClassification.from_pretrained('ShihTing/PanJuOffset_TwoClass')
# Predict
inputs = extractor(images=image, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits
Prob = outputs.logits.softmax(dim=-1).tolist()
print(Prob)
# model predicts one of the 1000 ImageNet classes
predicted_class_idx = logits.argmax(-1).item()
print("Predicted class:", model.config.id2label[predicted_class_idx])